Software system for automatic classification of cirrus cloud textures on the basis of neuro-fuzzy network

Download article in PDF format

Authors: Astafurov V. G., Evsyutkin T. V.

Annotation: The paper describes specialized software for classification of cirrus cloud subtypes using MODIS satellite imagery. A technique of classifier design on the basis of neuro-fuzzy network and set of informative texture features of cirrus cloud subtypes is discussed. The functionality and the field of application of the software system developed are described, the software allows four cirrus cloud subtypes to be classified according to the current meteorological standard and the probability of correct classification to be estimated as 0,57.

Keywords: software, classification, neural network, texture features, cirrus clouds